100mm 2.8L Macro IS as a portrait lens

[BrettS]

[...]

More interesting for this discussion would be full-frame, uncropped comparisons "zoomed with the feet" to simulate the actual reason one would choose the one lens over the other.

+1

 

[comsense]

Thanks for summary, skitron and trumpet. There is nothing called portrait focal length and dof. Companies define that for simplicity sake (for simple minded client) and keeping real life reproduction in mind. At least for me, art photography is not about that. Focal length as well as dof are just a tool. I hope painting forum does not argue that size 1 brush is better that 2 for masterpieces. But you are allowed to say that this is my favorite focal length for portraits.
However, I disagree on subtle difference point. If you are able to see the subtle differences, it can be like tiny sand particle in eye. For people who can't see them, world is bliss. Passion for perfection is engine of creation. Life and things are usually not dramatic.
Specifically, the dof at f/2 vs f/3.5 for 135 @ 10 ft is .19 vs .32. thats not subtle mathematically, about 1.7 times more. But I get yr point, dof difference is visually subtle (still important to few), f/2 to f/3.5 can be difference between photo and no photo in certain low light situations.

 

[Plamen]

Indeed, it'll be just about the same difference as between 135 f/2 and 135 f/3.5. Different? Yes. Dramatically different? No.

Judge for yourself, 135/4 vs. 135/2 (f/4 is closer to the truth than f/3.5, and I do not have a f/3.5 version of that picture on my computer anyway):

 

[Marsu42]

So, if you stand in the same place and take a shot with a 50 f/2 and 200 f/8 and crop the 50 to the same angle of view as the 200, you get, for the most part, an identical image.

Thanks for the explanation! Though I have to comment/ask if the bokeh is not also very dependent on the lens build, i.e. the design of the aperture (blade number and rounded or not) esp. with point highlights in the background?

 

[TrumpetPower!]

Indeed, it'll be just about the same difference as between 135 f/2 and 135 f/3.5. Different? Yes. Dramatically different? No.

Judge for yourself, 135/4 vs. 135/2 (f/4 is closer to the truth than f/3.5, and I do not have a f/3.5 version of that picture on my computer anyway):

The f/2 exposure is also about a stop darker. Just rendering the background that much darker also serves to significantly make it fade away -- you've essentially got a double whammy going on.

Normalize the exposure between the two images and the difference will still be there, but it won't be as dramatic as your example suggests.

Indeed, you can do wonders for your OOF backgrounds in post-processing just by selectively darkening them by a stop or two -- which you can conveniently do with the vignette tool in most raw processors. And, yes -- that's why vignetting isn't such a problem with wide-open fast lenses, as well as part of the explanation for why fast glass produces better backgrounds; darker is almost always better when it comes to backgrounds.

So, if you stand in the same place and take a shot with a 50 f/2 and 200 f/8 and crop the 50 to the same angle of view as the 200, you get, for the most part, an identical image.

Thanks for the explanation! Though I have to comment/ask if the bokeh is not also very dependent on the lens build, i.e. the design of the aperture (blade number and rounded or not) esp. with point highlights in the background?

Aperture design will certainly change the shape of the out-of-focus points. It won't change their size.

The subjective quality of a lens's bokeh (smooth, confusing, buttery, harsh, whatever) is basically all about the shape, not the size.

Cheers,

b&

 

[Plamen]

The f/2 exposure is also about a stop darker. Just rendering the background that much darker also serves to significantly make it fade away -- you've essentially got a double whammy going on.

Normalize the exposure between the two images and the difference will still be there, but it won't be as dramatic as your example suggests.

Here we go (it was 0.2 stop darker), 135/2 vs. 135/4. I just realized that those are images taken with a crop body. Anyway,

 

[TrumpetPower!]

The f/2 exposure is also about a stop darker. Just rendering the background that much darker also serves to significantly make it fade away -- you've essentially got a double whammy going on.

Normalize the exposure between the two images and the difference will still be there, but it won't be as dramatic as your example suggests.

Here we go (it was 0.2 stop darker), 135/2 vs. 135/4. I just realized that those are images taken with a crop body. Anyway,

It's good to have both sets, as it illustrates that, yes, the darker exposure makes the background fade away more. Not as much as the wider aperture to be sure, but in the same ballpark.

b&

 

[Plamen]

This is shot on FF now. Too cold outside, so I took two quick shots inside. The doll is about 1.5x the size of a human head. In the 135L shot, the doll is a bit larger - I was too lazy to use a tripod and frame carefully. The "model" here is too close for the 135L to really create a dramatic difference but it is still different enough.

The harsh light in the background is there on purpose. The flowers were very close to the doll, maybe 50cm or so. Developed in LR, no lens corrections. The 135L image got less exposure.

135/2

100/2.8

 

[RLPhoto]

The 135L looks better. No surprise.

 

[TrumpetPower!]

The 135L looks better. No surprise.

Depends on how much you're looking to isolate the subject and how much you're looking to put it in context. Each has its place, in other words.

b&

 

[Marsu42]

I was too lazy to use a tripod and frame carefully.

The framing is rather different, moving the 100L towards the doll would have resulted in more background blur. No doubt the 135L is much more bokehlicious(tm), but if you want to do a comparison it'd be done properly, or you can also save the upload time...

Depends on how much you're looking to isolate the subject and how much you're looking to put it in context. Each has its place, in other words.

You constantly write reasonable things, please keep posting on CR

 

[comsense]

The 135L looks better. No surprise.

Depends on how much you're looking to isolate the subject and how much you're looking to put it in context. Each has its place, in other words.

b&

Would you guys just give up?? Neither maths nor Bokeh is on your side
Argument was whether its clearly different, not whether one is desired over other
Quick maths I did:
DOF of 135 f/2 on FF @ 10 ft = .19 ft
DOF of 100 f/2.8 on FF @ 10 ft = .5 ft, 163% or 2.6 times more
And the equivalent f stop to get same DOF on 135 mm is f/5.6 not f/3.5 as you suggested.
And don't tell me f/5.6 to f/2 is just subtle, unless you have different definition for subtle.
You can't quantify bokeh so see plamen's post
It can't be more different and clear than this

 

[Plamen]

I was too lazy to use a tripod and frame carefully.

The framing is rather different, moving the 100L towards the doll would have resulted in more background blur. No doubt the 135L is much more bokehlicious(tm), but if you want to do a comparison it'd be done properly, or you can also save the upload time...

Wow - tough crowd! We have 12+ pages of posts and this is the first direct comparison, if I am not mistaken. And I am been hammered for 10% or so difference in framing, and for wasting bandwidth!

 

[Marsu42]

Wow - tough crowd! We have 12+ pages of posts and this is the first direct comparison, if I am not mistaken. And I am been hammered for 10% or so difference in framing, and for wasting bandwidth!

I didn't want to cause undue alarm or attack you, and the surgar-coated version is also available ... I just tried pointing out that some tiny headroom for even further future improvement exists, but your input is very much appreciated and even at the current state great and very helpful

Maybe I'm a bit over-critical, but I often observe researchers that for some matter discover exactly what they went out to find and knew before - it was a bit like that with your comparison shots: We know the 135L has better bokeh, so why bother with a real comparison to prove it ?

 

[Northstar]

let's try to find some common ground...I'm sure we can all agree that the canon 135L and 100L are both superior to the Nikon 135 and 105? :

also, you know what I would pay to see at this point...a simple bet with real money, I provide 10 random shots with the 135 at F2, and 10 shots with the 100 at 2.8. and RLphoto and Privatedesign put their money up....for each photo that rlphoto correctly identifies as the 135, he gets $100 from Private, and for each photo he get's wrong...he pays $100 to private.

if we could all gather around the monitor with our favorite beverage, that would be some good fun! ;D

 

[comsense]

Wow - tough crowd! We have 12+ pages of posts and this is the first direct comparison, if I am not mistaken. And I am been hammered for 10% or so difference in framing, and for wasting bandwidth!

I didn't want to cause undue alarm or attack you, and the surgar-coated version is also available ... I just tried pointing out that some tiny headroom for even further future improvement exists, but your input is very much appreciated and even at the current state great and very helpful

Maybe I'm a bit over-critical, but I often observe researchers that for some matter discover exactly what they went out to find and knew before - it was a bit like that with your comparison shots: We know the 135L has better bokeh, so why bother with a real comparison to prove it ?

Good you said that it was obvious. Now turn around and tell that to people who are hammering everyone who tries to tell this 'obvious' mathematical fact (as far as subject isolation is concerned; bokeh cant be quantified). They claim that difference is so subtle that most people cant pick it up in blind tests. Then, they either show crappy examples or 135 used at f/8 to match the DOF (dah, its going to be equal because you matched it) to make a point. Plamen shows first direct and relevant comparison and gets hammered by you for showing the 'obvious'. Fair world......

 

[RLPhoto]

Wow - tough crowd! We have 12+ pages of posts and this is the first direct comparison, if I am not mistaken. And I am been hammered for 10% or so difference in framing, and for wasting bandwidth!

I didn't want to cause undue alarm or attack you, and the surgar-coated version is also available ... I just tried pointing out that some tiny headroom for even further future improvement exists, but your input is very much appreciated and even at the current state great and very helpful

Maybe I'm a bit over-critical, but I often observe researchers that for some matter discover exactly what they went out to find and knew before - it was a bit like that with your comparison shots: We know the 135L has better bokeh, so why bother with a real comparison to prove it ?

Good you said that it was obvious. Now turn around and tell that to people who are hammering everyone who tries to tell this 'obvious' mathematical fact (as far as subject isolation is concerned; bokeh cant be quantified). They claim that difference is so subtle that most people cant pick it up in blind tests. Then, they either show crappy examples or 135 used at f/8 to match the DOF (dah, its going to be equal because you matched it) to make a point. Plamen shows first direct and relevant comparison and gets hammered by you for showing the 'obvious'. Fair world......

Or some simply just can't accept that "I told you so."

 

[RLPhoto]

Privatebydesign - number of unique images made. - 0

I don't believe you have a relevant opinion in determining character in comparing unique lenses.

 

[Plamen]

First, the "comparison" plamen posted IS NOT relevant or the situation I have constantly referred to, I KEEP saying FOR THE SAME FRAMED IMAGE, that means moving forwards for the 100.

Exactly what I did. The error was about 10%.

I will post some more.

 

[Plamen]

Why not just try and tell me which of the four images I posted many pages ago are shot with which lens, surely that should be easy seeing as how the 135 has a "unique look"? And that is my point, yet again, it is not about comparisons, it is about the FACT that nobody can RELIABLY tell what image was shot with which lens, nobody who alludes to this "unique look" can reliably identify it, if you can't reliably identify it it isn't "unique".

I answered that already, several times. It is not hard to use a lens with a unique look in a way that nobody can tell the difference with a f/4 zoom. It has unique look when you use it in a unique way.

I have comparisons of the 24-105 with the 50L, the 85LII, the 35L. You would not be able to tell a difference, and even under 100% it will be hard to say which is which. I can post them if you insist. So what now, since nobody can tell the difference in those shots, those three primes are useless?